This study (conducted between July-December 2017) used the CHNRI method (see Box 1 ). The invited experts generated and systematically ranked research questions (RQs) that could potentially assist Iran in achieving its long-term health targets, as outlined in the Iran’s General Health Policies (GHPs) and the United Nations’ Sustainable Development Goals (SDGs). The priorities were identified for the next five years to be aligned with Iran’s major national plans. Detailed information about different components of the context of this exercise and the process in which the context was defined is provided in Table S1 in Online Supplementary Document(Online Supplementary Document) , and a full list of the included health targets is available in Table S2 in Online Supplementary Document(Online Supplementary Document) .

To rank the RQs, “weighted research priority scores” were calculated. This score took into account the average score provided by the experts to each RQ across the 5 criteria, and the weights assigned by the stakeholders to each criterion. The average scores – that initially were between 0 and 1 – were multiplied by 100 to provide scores in a range between 0 and 100. The RQs were also ranked based on the scores for each criterion. Average expert agreement (AEA) – ie, the level of agreement among the scorers – on each RQ was also computed as the frequency of the mode (ie, the most common score divided by the total number of scores). All the independent responses of scorers, calculations that led to the final scores, weights, and AEA are provided in Online Supplementary Document(Online Supplementary Document), in a Microsoft Excel (Microsoft Inc, Seattle WA, USA) file.

All of the original materials and correspondences with participants in this exercise were in Persian, and the ones that required further discussion within the steering committee were translated into English by one of the authors (Parisa Mansoori). To ensure that the researchers felt free to express their opinions while generating and/or scoring RQs, they were informed that nobody outside the steering committee of the exercise would be able to link their input to their real names.

In total, 128 RQs were systematically scored by 48 experts against five criteria. Sixty-eight representatives from stakeholder groups provided weights to the five criteria (see Box 2 ). The weighted Research Priority Scores (wRPS) ranged from 84.5% to 28.5%. Table 2 shows the top 10 priorities, the wRPSs, the score that each priority has received for each criterion, and the AEA. Full list of the ranked 128 RQs is provided in Online Supplementary Document(Online Supplementary Document), Excel file. Ninety-two of the 128 RQs (71.8%) fell entirely or partially under Health Policy and Systems Research (HPSR – using the definition of the Alliance for Health Policy and Systems Research [15]). Forty-five RQs addressed specific causes of morbidity, mortality, and/or interventions for their diagnosis, prevention, and/or management. Sixteen RQs fell into the category of “Planetary Health” (ie, addressing human health effects of accelerating environmental changes – the term was coined in 2014 [16]). Other themes that were found across the 128 RQs were traditional Iranian medicine and Iranian-Islamic values (9 RQs); medical education, science, and innovation (6 RQs); and community participation (5 RQs).

The top 10 priorities comprised a wide range of RQs: 9 contained HPSR components with one RQ being focused on medical education; 6 entirely or partially fell into the scope of epidemiological research; and one addressed planetary health. The top priority RQ “How do we reform Iran's health insurance system to improve equity?” also scored highest for impact on economy (RPS = 90.7) and equity (RPS = 91). The second, the third, and the fifth priorities all addressed NCDs: two HPSR and one epidemiological question. They focused on NCD prevention strategy for Iran and its integration into the primary health care; seeking effective and cost-effective population-level interventions for reducing and managing NCDs and road traffic injuries; and an epidemiological assessment of NCDs and their underlying causes across Iran and making projections for 2030. Both of the HPSR questions that addressed NCDs have also scored highest for impact on health.

The 4^th priority focused on tailoring medical qualifications to better serve the needs of the nation, and it was agreed among the experts that this RQ would provide the greatest opportunity for capacity building. The experts collectively proposed that the 6^th priority, which focused on ways to achieve Universal Health Coverage (UHC), was identified as a rather feasible question to address (ranked 8^th for feasibility) and with a good potential for making an impact on health (ranked 4^th). This priority was followed by proposing investigation of the current and the future common health problems in Iran’s elderly population and identifying strategies to reduce their economic burden. The consensus of experts suggested that this question is the most feasible among all the 128 RQs.

One planetary health question that was identified as a top-10 priority addressed health impacts of water crisis in Iran and how the health system should respond to it. Another priority suggesting a health systems research to replace physician-centered system with a team-based one was found to provide a great opportunity for capacity building. Another top-10 priority proposed assessing the status of antibiotic resistance in Iran and investigating ways to promote rational use of antibiotics.

In general, the level of agreement among the experts was between 67.5% and 34.58%. The RQs for which the greatest level of expert agreement was observed were generally those that received the highest RPS. Seven of the top-10 priorities had the greatest agreement scores, with the top-3 RQs having had received the highest AEA (65%, 66%, and 68%). On the other hand, RQs that had been the most controversial were those that ranked lower than the median based on the expert scoring (RPS<68.3).

Among the 10 RQs from the bottom of the ranking (wRPS range = 52.5-28.4), 8 addressed traditional Iranian medicine and Islamic-Iranian values, while two were very specific questions related to dementia. Examples of low-scoring RQs related to dementia were “How do behavioural and psychiatric symptoms of dementia patients differ between Iran and other countries?” (wRPS = 41.4); or “At which stages of dementia do people seek medical assistance in Iran, and how does this vary between groups with different levels of education?” (wRPS = 52.5). The RQ that received the lowest score (wRPS = 28.4) was: “How should we remember and honour the most influential medical scientists in Iran's history?”

This exercise identified the research priorities that have a potential to assist Iran in addressing the knowledge gaps to achieve its long-term health targets as outlined in the National GHPs by 2025 and the SDGs by 2030. This was the first national-level health research priority-setting exercise in Iran that aggregated independent input from prominent Iranian researchers, policy makers and funders, and a group of stakeholders from the wider society. While the identified research priorities covered a diverse range of issues related to health, the majority of the proposed RQs, as well as 90% of the top-ranked questions, had aspects of HPSR. This is not surprising given the context of this study, which addressed a set of broad national health targets within a relatively short period (ie, the next 5 years) [11].

The HPSR questions that ranked among the top quartile of the 128 RQs predominantly related to financing, governance, and/or service delivery. This corresponds quite well with Iran’s ongoing Health Transformation Plan—launched in 2014—which aims to provide UHC by 2025, improve the quality of health care services, and achieve financial protection [17]. Furthermore, seeking ways to improve efficiency of the health system was reflected in several of the top-10 RQs, eg, reform of the insurance system; prevention and management of NCDs at different levels of the health system; and replacing the physician-centered health care system with a team-based one. This may suggest that, although Iran’s investment in health care has substantially increased since 2015 [17], the system has not yet succeeded in efficiently using its resources, either financial or human resources.

The priorities also clearly highlighted the significance of addressing questions that seek cost-effective population-level interventions and health system strategies to prevent NCDs and road traffic injuries in Iran. It was also agreed that studying the current and the estimated future distribution of NCDs and their underlying causes across Iran as well as the common health problems in Iran’s elderly population should be prioritized. Such results reflect the need for further investment in addressing the knowledge gaps in a country facing a rapidly increasing burden of NCDs [2] with an ageing population [18]. Such challenges reinforce the need for improving efficiency of the health system.

Consensus of the participants also indicated that Iran’s success in achieving its long-term health targets requires investment in studying the health impacts of environmental challenges, eg, water crisis. It had been argued previously that securing sustainable water resources in Iran requires immediate coordinated multi-sectorial action [19]. Another top-10 priority highlighted the importance of tailoring medical education to better serve the needs of the society. This question was raised for the primary qualification (eg, Doctor of Medicine) and was also specified for the curriculum of specialist training in psychiatry.

The questions related to the values embedded in the rich history of medical sciences in Iran were relatively prevalent among the originally proposed RQs (9 out of 128). However, the majority of them were ranked near the bottom of the list. This could suggest that, although a number of Iranian academics agreed that knowledge gaps regarding these issues should be addressed, it seems that it was difficult for these questions to meet the criteria of this exercise. It is likely that in a prioritization exercise within a different context, eg, considering a longer time scale, or using a different set of criteria, these questions could receive a higher average score.

It is worth mentioning that since 2000, Iran’s MOHME has been building capacity in medical universities across the country and in different departments of the Ministry to carry out research prioritization exercises [20]. Research priorities had been identified either at the national level in specific areas, eg, prevention of cervical cancer [21], HIV and AIDS [22], and patient safety [23], or at the institutional level, eg, in health systems research [24], or medical education [25]. However, none of these studies had engaged appropriate stakeholders from the wider society; the majority lacked a clearly-defined context; and all had used methods that included a panel of experts, rather than a structured and replicable approach.

Furthermore, while efforts to improve governance of health research in a developing country are admirable, it is unclear how the priorities were being set at the national level. It was reported in 2016 that a total of 6723 “research priorities” were collected from Iran’s medical universities [26], but this number is too large to consider each idea a “priority”. The report [26] continued that the collected research questions were grouped into nine “main areas”, but those were also too broad to be considered priorities that could direct national health research investments. A qualitative study investigated the barriers to evidence-based decision-making in Iran’s health system in 2012 [8]. The participants identified the following issues: (i) lack of a systematic prioritization mechanism; (ii) priorities being identified by MOHME and not being communicated with academics; and (iii) priorities being too general, thus failing to guide researchers.

Our study adapted the CHNRI method, which has several advantages ( Box 1 ) [10-13]. Compared to previous national-level CHNRI exercises conducted in South Africa, India, Brazil, China, and one study that included Malawi, Nigeria, and Zimbabwe [11,27], this study had a relatively similar time scale, ie, 5-10 years. Perhaps this shows that it is of more interest for national-level studies to set the priorities with shorter periods in mind. This study had several strengths compared to previous CHNRI exercises. As the method has already been applied in over 50 exercises [11], plenty of published guidance was available to help adapt the method to our needs. The management group considered the modifications in the previous CHNRI studies [11] and this helped to easily reach a consensus over the five criteria most relevant to the present study. Furthermore, previous CHNRI exercises had a response rate between 30-70% [11], while in this exercise more than 70% of contacted participants responded both to a request to generate RQs and to score them. This shows that the exercise has been successful in engaging the experts with the process and the results are less likely to be biased. Moreover, the wide range of final research priority scores (84.5-28.5%) indicated a good level of discrimination between the RQs in meeting the five criteria. Finally, only 24% of previous CHNRI exercises had managed to engage an appropriate group of stakeholders in the process [14].

We explored the impact that the stakeholders’ assigned weights had on the final ranking of the RQs. Their input led to minor changes in the order of the questions, and also promoted two new RQs among the top-10 priorities, one addressing the common health problems in the elderly and another regarding antibiotic resistance. In the top quartile, notable examples of questions that were helped by the stakeholders' input were: (i) impact of climate change on air quality of Iranian cities and its health effects; (ii) community-based strategies to reduce energy consumption and environmental pollutants to improve health; and (iii) community-based interventions for reducing high-risk behaviors (intravenous drug use, unprotected sex) amongst adolescents living in deprived areas. In summary, the stakeholders’ input did not lead to a drastic change in the final ranks, but it helped several RQs with a focus on community participation to improve their ranks.

In terms of the level of agreement among experts, the AEA ranged between 67.5 and 34.5%. All of the RQs that were the most controversial were RQs that have ranked lower than the median RPS of 68.3. The greatest level of controversy seems to have been observed among the questions that were either too broad or too specific. This study had some apparent limitations, summarized in Box 3 .

The identified priorities seem plausible within the context of the study, which involves a country with an ageing population, facing a high burden of NCDs and road traffic injuries while in parts still struggling with endemics of infectious diseases, experiencing accelerated environmental changes, having a large number of students at universities of medical sciences, and undergoing a Health Transformation Plan to achieve UHC. On the other hand, we should analyse why the majority of questions addressing mental health, substance abuse, or dementia did not score highly despite a high burden of disability and death which they cause in Iran [4]. It is possible that first, the way in which the questions in those areas were framed did not seem feasible enough to the scorers. Second, the low scores that these questions have received for impact on economy may suggest insufficient awareness of the majority of the engaged experts about the burden that non-physical health conditions could cause. It is also worth mentioning that while the consensus of experts in this exercise strongly supported the need for investment in further studies on NCDs, one of the top-10 priorities addressed antimicrobial resistance. This highlighted that although communicable diseases are contributing to less DALYs in Iran, they continue to remain important. Diseases such as influenza, brucellosis, and tuberculosis are still major endemics in Iran [3], and the country’s capacity in studying and controlling outbreaks of neglected, emerging, and/or re-emerging infectious diseases requires improvements [3].

The identified priorities described in this paper guide research that is likely to be feasible to be conducted within the next five years, and bear a higher impact on health and economy and lead to capacity building and equity. The authors hope that this paper will first contribute to the field of health research priority-setting as one of the most recent applications of the CHNRI method at a national level. Second, this was the first application of the CHNRI method in Iran and we hope that the findings of this study will be utilized in making funding decisions within the next five years. Advocacy of the findings could be done through holding meetings with decision-makers at different departments of MOHME, the Islamic Parliament Research Center, Academy of Medical Sciences, and research departments of major Iranian universities of medical sciences. Summaries of the results should become available online for open and easy access of all researchers and stakeholders. We also encourage different funding agencies to compute the weights for the criteria using the input from their own major stakeholders to adjust the priorities with their organizations’ targets.

Iran has got a substantial health research capacity [6] and proved interest in prioritizing research [20]. Now that the CHNRI method has been successfully introduced in Iran, we recommend its application to identify the research priorities in more specific domains, eg, for certain health conditions with a high burden in Iran, or for specific target populations. It is also recommended to conduct the CHNRI studies at a sub-national level. This would assist with identifying the priorities based on the needs of each region of this relatively large and populous country, where each part deals with particular health challenges ranging from water scarcity and air pollution to substance abuse or challenges imposed by sharing borders with countries in conflict. The development of a Massive Open Online Course (MOOC) for the Iranian audience could facilitate implementation of the method at a larger scale. We also recommend this national-level exercise to be repeated every five years to be aligned with Iran’s 5-Year Development Plans. Within the next five years, we will be tracking research grant calls and application to monitor if the findings of this study are implemented, and will highlight the priorities that will not be addressed. Finally, this exercise was the first implication of the CHNRI method in the World Health Organization (WHO) Eastern Mediterranean Region (EMR). The research team will pursue efforts to share the results with the WHO EMR Office that may assist conduct of the exercise in countries with a fairly similar context to Iran.

The main messages of this study were that addressing equity by reforming insurance system and improving equality in health services access should be prioritized in Iran. Furthermore, research on preventive strategies for the leading NCDs and road traffic injuries are needed, as well as further epidemiological research on NCDs and the health problems of the elderly population. This should happen in parallel to a tailored curriculum of qualifications in the universities of medical sciences, where a team-based health care system should be promoted. Research is also needed on over-/mis-use of antibiotics and the health impacts of water crisis.

This study was a good example of a national-level health research priority-setting process and a rare example where a national research funding body used the process to decide on health research funding in a holistic way. This helped to overcome problems of previous research priority-setting exercises, with further implementation now under way in Iran. This exercise also provides a model to public research funding bodies in other countries that could be adapted to improve decision-making on funding allocation for health research, especially in low resource settings.